manugandham27

Multi Agent LLM Council — AI skill for Claude Code

AI community

Multi-LLM consensus & evaluation platform querying OpenAI, Gemini, Claude, DeepSeek & Mistral in parallel to calculate weighted consensus, measure vote entropy, and eliminate single-model hallucinatio.

How to install Multi Agent LLM Council

This entry records only its repository, not the path inside it, so there is no exact command to give. Open manugandham27/Multi-Agent-LLM-Council and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Multi Agent LLM Council does

Alternatives in AI

Define — Definition phase - Clarify and scope problems with multi-AI consensus 2.8k ★
VibeMeter — A project to measure costs for Cursor and other AI provider 379 ★
Claudecode Harness — Production-grade orchestration harness for Claude Code — Context discipline, subagents, and multi-model consen 223 ★

README

🏛️ Multi-Agent LLM Council

Ask multiple AI models. Compare their reasoning. Reach a transparent consensus.

[![Next.js](https://img.shields.io/badge/Frontend-Next.js%2014-000000?style=for-the-badge&logo=nextdotjs&logoColor=white)](https://nextjs.org/) [![Vercel](https://img.shields.io/badge/Vercel-Live_Production-000000?style=for-the-badge&logo=vercel&logoColor=white)](https://web-psi-six-97.vercel.app) [![NestJS](https://img.shields.io/badge/Backend-NestJS-E0234E?style=for-the-badge&logo=nestjs&logoColor=white)](https://nestjs.com/) [![FastAPI](https://img.shields.io/badge/AI_Service-FastAPI-009688?style=for-the-badge&logo=fastapi&logoColor=white)](https://fastapi.tiangolo.com/) [![Prisma](https://img.shields.io/badge/ORM-Prisma-2D3748?style=for-the-badge&logo=prisma&logoColor=white)](https://www.prisma.io/) [![License](https://img.shields.io/badge/License-MIT-blue?style=for-the-badge)](LICENSE)

[Live Demo](https://web-psi-six-97.vercel.app) • [Features](#-key-features) • [Architecture](#-system-architecture) • [Quickstart](#-quickstart-guide) • [Consensus Methodology](#-consensus-engine)

[!TIP] **🌐 Live Production Web Application**: Experience the deployed application live on Vercel at **[https://web-psi-six-97.vercel.app](https://web-psi-six-97.vercel.app)**!

[!IMPORTANT] **Multi-Agent LLM Council** is an enterprise-grade AI ensemble platform that queries multiple leading LLM providers (OpenAI GPT, Google Gemini, Anthropic Claude, DeepSeek, Mistral, and local Ollama models) in parallel, measures vote entropy/disagreement, executes weighted consensus algorithms, and invokes an independent Judge verifier to eliminate single-model hallucinations.

📖 About The Project

**Multi-Agent LLM Council** was built to solve the fundamental reliability problem in modern Artificial Intelligence: **single-model hallucinations and domain bias**.

By orchestrating an ensemble of competing LLM models in parallel, th